1ee3c74fbSChris Lattner //===- FileCheck.cpp - Check that File's Contents match what is expected --===//
2ee3c74fbSChris Lattner //
3ee3c74fbSChris Lattner //                     The LLVM Compiler Infrastructure
4ee3c74fbSChris Lattner //
5ee3c74fbSChris Lattner // This file is distributed under the University of Illinois Open Source
6ee3c74fbSChris Lattner // License. See LICENSE.TXT for details.
7ee3c74fbSChris Lattner //
8ee3c74fbSChris Lattner //===----------------------------------------------------------------------===//
9ee3c74fbSChris Lattner //
10ee3c74fbSChris Lattner // FileCheck does a line-by line check of a file that validates whether it
11ee3c74fbSChris Lattner // contains the expected content.  This is useful for regression tests etc.
12ee3c74fbSChris Lattner //
13ee3c74fbSChris Lattner // This program exits with an error status of 2 on error, exit status of 0 if
14ee3c74fbSChris Lattner // the file matched the expected contents, and exit status of 1 if it did not
15ee3c74fbSChris Lattner // contain the expected contents.
16ee3c74fbSChris Lattner //
17ee3c74fbSChris Lattner //===----------------------------------------------------------------------===//
18ee3c74fbSChris Lattner 
1939a0ffc3SMichael J. Spencer #include "llvm/ADT/OwningPtr.h"
2091d19d8eSChandler Carruth #include "llvm/ADT/SmallString.h"
2191d19d8eSChandler Carruth #include "llvm/ADT/StringExtras.h"
2291d19d8eSChandler Carruth #include "llvm/ADT/StringMap.h"
23ee3c74fbSChris Lattner #include "llvm/Support/CommandLine.h"
24ee3c74fbSChris Lattner #include "llvm/Support/MemoryBuffer.h"
25ee3c74fbSChris Lattner #include "llvm/Support/PrettyStackTrace.h"
26f08d2db9SChris Lattner #include "llvm/Support/Regex.h"
2791d19d8eSChandler Carruth #include "llvm/Support/Signals.h"
28ee3c74fbSChris Lattner #include "llvm/Support/SourceMgr.h"
29ee3c74fbSChris Lattner #include "llvm/Support/raw_ostream.h"
307b6fef82SMichael J. Spencer #include "llvm/Support/system_error.h"
318879e06dSChris Lattner #include <algorithm>
32e8b8f1bcSEli Bendersky #include <map>
33e8b8f1bcSEli Bendersky #include <string>
34e8b8f1bcSEli Bendersky #include <vector>
35ee3c74fbSChris Lattner using namespace llvm;
36ee3c74fbSChris Lattner 
37ee3c74fbSChris Lattner static cl::opt<std::string>
38ee3c74fbSChris Lattner CheckFilename(cl::Positional, cl::desc("<check-file>"), cl::Required);
39ee3c74fbSChris Lattner 
40ee3c74fbSChris Lattner static cl::opt<std::string>
41ee3c74fbSChris Lattner InputFilename("input-file", cl::desc("File to check (defaults to stdin)"),
42ee3c74fbSChris Lattner               cl::init("-"), cl::value_desc("filename"));
43ee3c74fbSChris Lattner 
44ee3c74fbSChris Lattner static cl::opt<std::string>
45ee3c74fbSChris Lattner CheckPrefix("check-prefix", cl::init("CHECK"),
46ee3c74fbSChris Lattner             cl::desc("Prefix to use from check file (defaults to 'CHECK')"));
47ee3c74fbSChris Lattner 
482c3e5cdfSChris Lattner static cl::opt<bool>
492c3e5cdfSChris Lattner NoCanonicalizeWhiteSpace("strict-whitespace",
502c3e5cdfSChris Lattner               cl::desc("Do not treat all horizontal whitespace as equivalent"));
512c3e5cdfSChris Lattner 
5274d50731SChris Lattner //===----------------------------------------------------------------------===//
5374d50731SChris Lattner // Pattern Handling Code.
5474d50731SChris Lattner //===----------------------------------------------------------------------===//
5574d50731SChris Lattner 
563b40b445SChris Lattner class Pattern {
570a4c44bdSChris Lattner   SMLoc PatternLoc;
580a4c44bdSChris Lattner 
59eba55822SJakob Stoklund Olesen   /// MatchEOF - When set, this pattern only matches the end of file. This is
60eba55822SJakob Stoklund Olesen   /// used for trailing CHECK-NOTs.
61eba55822SJakob Stoklund Olesen   bool MatchEOF;
62eba55822SJakob Stoklund Olesen 
6391a1b2c9SMichael Liao   /// MatchNot
6491a1b2c9SMichael Liao   bool MatchNot;
6591a1b2c9SMichael Liao 
6691a1b2c9SMichael Liao   /// MatchDag
6791a1b2c9SMichael Liao   bool MatchDag;
6891a1b2c9SMichael Liao 
69b16ab0c4SChris Lattner   /// FixedStr - If non-empty, this pattern is a fixed string match with the
70b16ab0c4SChris Lattner   /// specified fixed string.
71221460e0SChris Lattner   StringRef FixedStr;
72b16ab0c4SChris Lattner 
73b16ab0c4SChris Lattner   /// RegEx - If non-empty, this is a regex pattern.
74b16ab0c4SChris Lattner   std::string RegExStr;
758879e06dSChris Lattner 
7692987fb3SAlexander Kornienko   /// \brief Contains the number of line this pattern is in.
7792987fb3SAlexander Kornienko   unsigned LineNumber;
7892987fb3SAlexander Kornienko 
798879e06dSChris Lattner   /// VariableUses - Entries in this vector map to uses of a variable in the
808879e06dSChris Lattner   /// pattern, e.g. "foo[[bar]]baz".  In this case, the RegExStr will contain
818879e06dSChris Lattner   /// "foobaz" and we'll get an entry in this vector that tells us to insert the
828879e06dSChris Lattner   /// value of bar at offset 3.
838879e06dSChris Lattner   std::vector<std::pair<StringRef, unsigned> > VariableUses;
848879e06dSChris Lattner 
85e8b8f1bcSEli Bendersky   /// VariableDefs - Maps definitions of variables to their parenthesized
86e8b8f1bcSEli Bendersky   /// capture numbers.
87e8b8f1bcSEli Bendersky   /// E.g. for the pattern "foo[[bar:.*]]baz", VariableDefs will map "bar" to 1.
88e8b8f1bcSEli Bendersky   std::map<StringRef, unsigned> VariableDefs;
898879e06dSChris Lattner 
903b40b445SChris Lattner public:
913b40b445SChris Lattner 
9291a1b2c9SMichael Liao   Pattern(bool matchEOF = false)
9391a1b2c9SMichael Liao     : MatchEOF(matchEOF), MatchNot(false), MatchDag(false) { }
9474d50731SChris Lattner 
950b707eb8SMichael Liao   /// getLoc - Return the location in source code.
960b707eb8SMichael Liao   SMLoc getLoc() const { return PatternLoc; }
970b707eb8SMichael Liao 
9843d50d4aSEli Bendersky   /// ParsePattern - Parse the given string into the Pattern.  SM provides the
9943d50d4aSEli Bendersky   /// SourceMgr used for error reports, and LineNumber is the line number in
10043d50d4aSEli Bendersky   /// the input file from which the pattern string was read.
10143d50d4aSEli Bendersky   /// Returns true in case of an error, false otherwise.
10292987fb3SAlexander Kornienko   bool ParsePattern(StringRef PatternStr, SourceMgr &SM, unsigned LineNumber);
1033b40b445SChris Lattner 
1043b40b445SChris Lattner   /// Match - Match the pattern string against the input buffer Buffer.  This
1053b40b445SChris Lattner   /// returns the position that is matched or npos if there is no match.  If
1063b40b445SChris Lattner   /// there is a match, the size of the matched string is returned in MatchLen.
1078879e06dSChris Lattner   ///
1088879e06dSChris Lattner   /// The VariableTable StringMap provides the current values of filecheck
1098879e06dSChris Lattner   /// variables and is updated if this match defines new values.
1108879e06dSChris Lattner   size_t Match(StringRef Buffer, size_t &MatchLen,
1118879e06dSChris Lattner                StringMap<StringRef> &VariableTable) const;
112b16ab0c4SChris Lattner 
113e0ef65abSDaniel Dunbar   /// PrintFailureInfo - Print additional information about a failure to match
114e0ef65abSDaniel Dunbar   /// involving this pattern.
115e0ef65abSDaniel Dunbar   void PrintFailureInfo(const SourceMgr &SM, StringRef Buffer,
116e0ef65abSDaniel Dunbar                         const StringMap<StringRef> &VariableTable) const;
117e0ef65abSDaniel Dunbar 
118*f8bd2e5bSStephen Lin   bool hasVariable() const { return !(VariableUses.empty() &&
119*f8bd2e5bSStephen Lin                                       VariableDefs.empty()); }
120*f8bd2e5bSStephen Lin 
12191a1b2c9SMichael Liao   void setMatchNot(bool Not) { MatchNot = Not; }
12291a1b2c9SMichael Liao   bool getMatchNot() const { return MatchNot; }
12391a1b2c9SMichael Liao 
12491a1b2c9SMichael Liao   void setMatchDag(bool Dag) { MatchDag = Dag; }
12591a1b2c9SMichael Liao   bool getMatchDag() const { return MatchDag; }
12691a1b2c9SMichael Liao 
127b16ab0c4SChris Lattner private:
1288879e06dSChris Lattner   static void AddFixedStringToRegEx(StringRef FixedStr, std::string &TheStr);
129e8b8f1bcSEli Bendersky   bool AddRegExToRegEx(StringRef RS, unsigned &CurParen, SourceMgr &SM);
130e8b8f1bcSEli Bendersky   void AddBackrefToRegEx(unsigned BackrefNum);
131fd29d886SDaniel Dunbar 
132fd29d886SDaniel Dunbar   /// ComputeMatchDistance - Compute an arbitrary estimate for the quality of
133fd29d886SDaniel Dunbar   /// matching this pattern at the start of \arg Buffer; a distance of zero
134fd29d886SDaniel Dunbar   /// should correspond to a perfect match.
135fd29d886SDaniel Dunbar   unsigned ComputeMatchDistance(StringRef Buffer,
136fd29d886SDaniel Dunbar                                const StringMap<StringRef> &VariableTable) const;
13792987fb3SAlexander Kornienko 
13892987fb3SAlexander Kornienko   /// \brief Evaluates expression and stores the result to \p Value.
13992987fb3SAlexander Kornienko   /// \return true on success. false when the expression has invalid syntax.
14092987fb3SAlexander Kornienko   bool EvaluateExpression(StringRef Expr, std::string &Value) const;
141061d2baaSEli Bendersky 
142061d2baaSEli Bendersky   /// \brief Finds the closing sequence of a regex variable usage or
143061d2baaSEli Bendersky   /// definition. Str has to point in the beginning of the definition
144061d2baaSEli Bendersky   /// (right after the opening sequence).
145061d2baaSEli Bendersky   /// \return offset of the closing sequence within Str, or npos if it was not
146061d2baaSEli Bendersky   /// found.
147061d2baaSEli Bendersky   size_t FindRegexVarEnd(StringRef Str);
1483b40b445SChris Lattner };
1493b40b445SChris Lattner 
1508879e06dSChris Lattner 
15192987fb3SAlexander Kornienko bool Pattern::ParsePattern(StringRef PatternStr, SourceMgr &SM,
15292987fb3SAlexander Kornienko                            unsigned LineNumber) {
15392987fb3SAlexander Kornienko   this->LineNumber = LineNumber;
1540a4c44bdSChris Lattner   PatternLoc = SMLoc::getFromPointer(PatternStr.data());
1550a4c44bdSChris Lattner 
15674d50731SChris Lattner   // Ignore trailing whitespace.
15774d50731SChris Lattner   while (!PatternStr.empty() &&
15874d50731SChris Lattner          (PatternStr.back() == ' ' || PatternStr.back() == '\t'))
15974d50731SChris Lattner     PatternStr = PatternStr.substr(0, PatternStr.size()-1);
16074d50731SChris Lattner 
16174d50731SChris Lattner   // Check that there is something on the line.
16274d50731SChris Lattner   if (PatternStr.empty()) {
16303b80a40SChris Lattner     SM.PrintMessage(PatternLoc, SourceMgr::DK_Error,
16403b80a40SChris Lattner                     "found empty check string with prefix '" +
16503b80a40SChris Lattner                     CheckPrefix+":'");
16674d50731SChris Lattner     return true;
16774d50731SChris Lattner   }
16874d50731SChris Lattner 
169221460e0SChris Lattner   // Check to see if this is a fixed string, or if it has regex pieces.
170d9466967STed Kremenek   if (PatternStr.size() < 2 ||
1718879e06dSChris Lattner       (PatternStr.find("{{") == StringRef::npos &&
1728879e06dSChris Lattner        PatternStr.find("[[") == StringRef::npos)) {
173221460e0SChris Lattner     FixedStr = PatternStr;
174221460e0SChris Lattner     return false;
175221460e0SChris Lattner   }
176221460e0SChris Lattner 
1778879e06dSChris Lattner   // Paren value #0 is for the fully matched string.  Any new parenthesized
17853e0679dSChris Lattner   // values add from there.
1798879e06dSChris Lattner   unsigned CurParen = 1;
1808879e06dSChris Lattner 
181b16ab0c4SChris Lattner   // Otherwise, there is at least one regex piece.  Build up the regex pattern
182b16ab0c4SChris Lattner   // by escaping scary characters in fixed strings, building up one big regex.
183f08d2db9SChris Lattner   while (!PatternStr.empty()) {
1848879e06dSChris Lattner     // RegEx matches.
18553e0679dSChris Lattner     if (PatternStr.startswith("{{")) {
18643d50d4aSEli Bendersky       // This is the start of a regex match.  Scan for the }}.
187f08d2db9SChris Lattner       size_t End = PatternStr.find("}}");
188f08d2db9SChris Lattner       if (End == StringRef::npos) {
189f08d2db9SChris Lattner         SM.PrintMessage(SMLoc::getFromPointer(PatternStr.data()),
19003b80a40SChris Lattner                         SourceMgr::DK_Error,
19103b80a40SChris Lattner                         "found start of regex string with no end '}}'");
192f08d2db9SChris Lattner         return true;
193f08d2db9SChris Lattner       }
194f08d2db9SChris Lattner 
195e53c95f1SChris Lattner       // Enclose {{}} patterns in parens just like [[]] even though we're not
196e53c95f1SChris Lattner       // capturing the result for any purpose.  This is required in case the
197e53c95f1SChris Lattner       // expression contains an alternation like: CHECK:  abc{{x|z}}def.  We
198e53c95f1SChris Lattner       // want this to turn into: "abc(x|z)def" not "abcx|zdef".
199e53c95f1SChris Lattner       RegExStr += '(';
200e53c95f1SChris Lattner       ++CurParen;
201e53c95f1SChris Lattner 
2028879e06dSChris Lattner       if (AddRegExToRegEx(PatternStr.substr(2, End-2), CurParen, SM))
2038879e06dSChris Lattner         return true;
204e53c95f1SChris Lattner       RegExStr += ')';
20553e0679dSChris Lattner 
2068879e06dSChris Lattner       PatternStr = PatternStr.substr(End+2);
2078879e06dSChris Lattner       continue;
2088879e06dSChris Lattner     }
2098879e06dSChris Lattner 
2108879e06dSChris Lattner     // Named RegEx matches.  These are of two forms: [[foo:.*]] which matches .*
2118879e06dSChris Lattner     // (or some other regex) and assigns it to the FileCheck variable 'foo'. The
2128879e06dSChris Lattner     // second form is [[foo]] which is a reference to foo.  The variable name
21357cb733bSDaniel Dunbar     // itself must be of the form "[a-zA-Z_][0-9a-zA-Z_]*", otherwise we reject
2148879e06dSChris Lattner     // it.  This is to catch some common errors.
21553e0679dSChris Lattner     if (PatternStr.startswith("[[")) {
216061d2baaSEli Bendersky       // Find the closing bracket pair ending the match.  End is going to be an
217061d2baaSEli Bendersky       // offset relative to the beginning of the match string.
218061d2baaSEli Bendersky       size_t End = FindRegexVarEnd(PatternStr.substr(2));
219061d2baaSEli Bendersky 
2208879e06dSChris Lattner       if (End == StringRef::npos) {
2218879e06dSChris Lattner         SM.PrintMessage(SMLoc::getFromPointer(PatternStr.data()),
22203b80a40SChris Lattner                         SourceMgr::DK_Error,
22303b80a40SChris Lattner                         "invalid named regex reference, no ]] found");
224f08d2db9SChris Lattner         return true;
225f08d2db9SChris Lattner       }
226f08d2db9SChris Lattner 
227061d2baaSEli Bendersky       StringRef MatchStr = PatternStr.substr(2, End);
228061d2baaSEli Bendersky       PatternStr = PatternStr.substr(End+4);
2298879e06dSChris Lattner 
2308879e06dSChris Lattner       // Get the regex name (e.g. "foo").
2318879e06dSChris Lattner       size_t NameEnd = MatchStr.find(':');
2328879e06dSChris Lattner       StringRef Name = MatchStr.substr(0, NameEnd);
2338879e06dSChris Lattner 
2348879e06dSChris Lattner       if (Name.empty()) {
23503b80a40SChris Lattner         SM.PrintMessage(SMLoc::getFromPointer(Name.data()), SourceMgr::DK_Error,
23603b80a40SChris Lattner                         "invalid name in named regex: empty name");
2378879e06dSChris Lattner         return true;
2388879e06dSChris Lattner       }
2398879e06dSChris Lattner 
24092987fb3SAlexander Kornienko       // Verify that the name/expression is well formed. FileCheck currently
24192987fb3SAlexander Kornienko       // supports @LINE, @LINE+number, @LINE-number expressions. The check here
24292987fb3SAlexander Kornienko       // is relaxed, more strict check is performed in \c EvaluateExpression.
24392987fb3SAlexander Kornienko       bool IsExpression = false;
24492987fb3SAlexander Kornienko       for (unsigned i = 0, e = Name.size(); i != e; ++i) {
24592987fb3SAlexander Kornienko         if (i == 0 && Name[i] == '@') {
24692987fb3SAlexander Kornienko           if (NameEnd != StringRef::npos) {
24792987fb3SAlexander Kornienko             SM.PrintMessage(SMLoc::getFromPointer(Name.data()),
24892987fb3SAlexander Kornienko                             SourceMgr::DK_Error,
24992987fb3SAlexander Kornienko                             "invalid name in named regex definition");
25092987fb3SAlexander Kornienko             return true;
25192987fb3SAlexander Kornienko           }
25292987fb3SAlexander Kornienko           IsExpression = true;
25392987fb3SAlexander Kornienko           continue;
25492987fb3SAlexander Kornienko         }
25592987fb3SAlexander Kornienko         if (Name[i] != '_' && !isalnum(Name[i]) &&
25692987fb3SAlexander Kornienko             (!IsExpression || (Name[i] != '+' && Name[i] != '-'))) {
2578879e06dSChris Lattner           SM.PrintMessage(SMLoc::getFromPointer(Name.data()+i),
25803b80a40SChris Lattner                           SourceMgr::DK_Error, "invalid name in named regex");
2598879e06dSChris Lattner           return true;
2608879e06dSChris Lattner         }
26192987fb3SAlexander Kornienko       }
2628879e06dSChris Lattner 
2638879e06dSChris Lattner       // Name can't start with a digit.
26483c74e9fSGuy Benyei       if (isdigit(static_cast<unsigned char>(Name[0]))) {
26503b80a40SChris Lattner         SM.PrintMessage(SMLoc::getFromPointer(Name.data()), SourceMgr::DK_Error,
26603b80a40SChris Lattner                         "invalid name in named regex");
2678879e06dSChris Lattner         return true;
2688879e06dSChris Lattner       }
2698879e06dSChris Lattner 
2708879e06dSChris Lattner       // Handle [[foo]].
2718879e06dSChris Lattner       if (NameEnd == StringRef::npos) {
272e8b8f1bcSEli Bendersky         // Handle variables that were defined earlier on the same line by
273e8b8f1bcSEli Bendersky         // emitting a backreference.
274e8b8f1bcSEli Bendersky         if (VariableDefs.find(Name) != VariableDefs.end()) {
275e8b8f1bcSEli Bendersky           unsigned VarParenNum = VariableDefs[Name];
276e8b8f1bcSEli Bendersky           if (VarParenNum < 1 || VarParenNum > 9) {
277e8b8f1bcSEli Bendersky             SM.PrintMessage(SMLoc::getFromPointer(Name.data()),
278e8b8f1bcSEli Bendersky                             SourceMgr::DK_Error,
279e8b8f1bcSEli Bendersky                             "Can't back-reference more than 9 variables");
280e8b8f1bcSEli Bendersky             return true;
281e8b8f1bcSEli Bendersky           }
282e8b8f1bcSEli Bendersky           AddBackrefToRegEx(VarParenNum);
283e8b8f1bcSEli Bendersky         } else {
2848879e06dSChris Lattner           VariableUses.push_back(std::make_pair(Name, RegExStr.size()));
285e8b8f1bcSEli Bendersky         }
2868879e06dSChris Lattner         continue;
2878879e06dSChris Lattner       }
2888879e06dSChris Lattner 
2898879e06dSChris Lattner       // Handle [[foo:.*]].
290e8b8f1bcSEli Bendersky       VariableDefs[Name] = CurParen;
2918879e06dSChris Lattner       RegExStr += '(';
2928879e06dSChris Lattner       ++CurParen;
2938879e06dSChris Lattner 
2948879e06dSChris Lattner       if (AddRegExToRegEx(MatchStr.substr(NameEnd+1), CurParen, SM))
2958879e06dSChris Lattner         return true;
2968879e06dSChris Lattner 
2978879e06dSChris Lattner       RegExStr += ')';
2988879e06dSChris Lattner     }
2998879e06dSChris Lattner 
3008879e06dSChris Lattner     // Handle fixed string matches.
3018879e06dSChris Lattner     // Find the end, which is the start of the next regex.
3028879e06dSChris Lattner     size_t FixedMatchEnd = PatternStr.find("{{");
3038879e06dSChris Lattner     FixedMatchEnd = std::min(FixedMatchEnd, PatternStr.find("[["));
3048879e06dSChris Lattner     AddFixedStringToRegEx(PatternStr.substr(0, FixedMatchEnd), RegExStr);
3058879e06dSChris Lattner     PatternStr = PatternStr.substr(FixedMatchEnd);
306f08d2db9SChris Lattner   }
307f08d2db9SChris Lattner 
30874d50731SChris Lattner   return false;
30974d50731SChris Lattner }
31074d50731SChris Lattner 
3118879e06dSChris Lattner void Pattern::AddFixedStringToRegEx(StringRef FixedStr, std::string &TheStr) {
312b16ab0c4SChris Lattner   // Add the characters from FixedStr to the regex, escaping as needed.  This
313b16ab0c4SChris Lattner   // avoids "leaning toothpicks" in common patterns.
314b16ab0c4SChris Lattner   for (unsigned i = 0, e = FixedStr.size(); i != e; ++i) {
315b16ab0c4SChris Lattner     switch (FixedStr[i]) {
316b16ab0c4SChris Lattner     // These are the special characters matched in "p_ere_exp".
317b16ab0c4SChris Lattner     case '(':
318b16ab0c4SChris Lattner     case ')':
319b16ab0c4SChris Lattner     case '^':
320b16ab0c4SChris Lattner     case '$':
321b16ab0c4SChris Lattner     case '|':
322b16ab0c4SChris Lattner     case '*':
323b16ab0c4SChris Lattner     case '+':
324b16ab0c4SChris Lattner     case '?':
325b16ab0c4SChris Lattner     case '.':
326b16ab0c4SChris Lattner     case '[':
327b16ab0c4SChris Lattner     case '\\':
328b16ab0c4SChris Lattner     case '{':
3298879e06dSChris Lattner       TheStr += '\\';
330b16ab0c4SChris Lattner       // FALL THROUGH.
331b16ab0c4SChris Lattner     default:
3328879e06dSChris Lattner       TheStr += FixedStr[i];
333b16ab0c4SChris Lattner       break;
334b16ab0c4SChris Lattner     }
335b16ab0c4SChris Lattner   }
336b16ab0c4SChris Lattner }
337b16ab0c4SChris Lattner 
338e8b8f1bcSEli Bendersky bool Pattern::AddRegExToRegEx(StringRef RS, unsigned &CurParen,
3398879e06dSChris Lattner                               SourceMgr &SM) {
340e8b8f1bcSEli Bendersky   Regex R(RS);
3418879e06dSChris Lattner   std::string Error;
3428879e06dSChris Lattner   if (!R.isValid(Error)) {
343e8b8f1bcSEli Bendersky     SM.PrintMessage(SMLoc::getFromPointer(RS.data()), SourceMgr::DK_Error,
34403b80a40SChris Lattner                     "invalid regex: " + Error);
3458879e06dSChris Lattner     return true;
3468879e06dSChris Lattner   }
3478879e06dSChris Lattner 
348e8b8f1bcSEli Bendersky   RegExStr += RS.str();
3498879e06dSChris Lattner   CurParen += R.getNumMatches();
3508879e06dSChris Lattner   return false;
3518879e06dSChris Lattner }
352b16ab0c4SChris Lattner 
353e8b8f1bcSEli Bendersky void Pattern::AddBackrefToRegEx(unsigned BackrefNum) {
354e8b8f1bcSEli Bendersky   assert(BackrefNum >= 1 && BackrefNum <= 9 && "Invalid backref number");
355e8b8f1bcSEli Bendersky   std::string Backref = std::string("\\") +
356e8b8f1bcSEli Bendersky                         std::string(1, '0' + BackrefNum);
357e8b8f1bcSEli Bendersky   RegExStr += Backref;
358e8b8f1bcSEli Bendersky }
359e8b8f1bcSEli Bendersky 
36092987fb3SAlexander Kornienko bool Pattern::EvaluateExpression(StringRef Expr, std::string &Value) const {
36192987fb3SAlexander Kornienko   // The only supported expression is @LINE([\+-]\d+)?
36292987fb3SAlexander Kornienko   if (!Expr.startswith("@LINE"))
36392987fb3SAlexander Kornienko     return false;
36492987fb3SAlexander Kornienko   Expr = Expr.substr(StringRef("@LINE").size());
36592987fb3SAlexander Kornienko   int Offset = 0;
36692987fb3SAlexander Kornienko   if (!Expr.empty()) {
36792987fb3SAlexander Kornienko     if (Expr[0] == '+')
36892987fb3SAlexander Kornienko       Expr = Expr.substr(1);
36992987fb3SAlexander Kornienko     else if (Expr[0] != '-')
37092987fb3SAlexander Kornienko       return false;
37192987fb3SAlexander Kornienko     if (Expr.getAsInteger(10, Offset))
37292987fb3SAlexander Kornienko       return false;
37392987fb3SAlexander Kornienko   }
37492987fb3SAlexander Kornienko   Value = llvm::itostr(LineNumber + Offset);
37592987fb3SAlexander Kornienko   return true;
37692987fb3SAlexander Kornienko }
37792987fb3SAlexander Kornienko 
378f08d2db9SChris Lattner /// Match - Match the pattern string against the input buffer Buffer.  This
379f08d2db9SChris Lattner /// returns the position that is matched or npos if there is no match.  If
380f08d2db9SChris Lattner /// there is a match, the size of the matched string is returned in MatchLen.
3818879e06dSChris Lattner size_t Pattern::Match(StringRef Buffer, size_t &MatchLen,
3828879e06dSChris Lattner                       StringMap<StringRef> &VariableTable) const {
383eba55822SJakob Stoklund Olesen   // If this is the EOF pattern, match it immediately.
384eba55822SJakob Stoklund Olesen   if (MatchEOF) {
385eba55822SJakob Stoklund Olesen     MatchLen = 0;
386eba55822SJakob Stoklund Olesen     return Buffer.size();
387eba55822SJakob Stoklund Olesen   }
388eba55822SJakob Stoklund Olesen 
389221460e0SChris Lattner   // If this is a fixed string pattern, just match it now.
390221460e0SChris Lattner   if (!FixedStr.empty()) {
391221460e0SChris Lattner     MatchLen = FixedStr.size();
392221460e0SChris Lattner     return Buffer.find(FixedStr);
393221460e0SChris Lattner   }
394221460e0SChris Lattner 
395b16ab0c4SChris Lattner   // Regex match.
3968879e06dSChris Lattner 
3978879e06dSChris Lattner   // If there are variable uses, we need to create a temporary string with the
3988879e06dSChris Lattner   // actual value.
3998879e06dSChris Lattner   StringRef RegExToMatch = RegExStr;
4008879e06dSChris Lattner   std::string TmpStr;
4018879e06dSChris Lattner   if (!VariableUses.empty()) {
4028879e06dSChris Lattner     TmpStr = RegExStr;
4038879e06dSChris Lattner 
4048879e06dSChris Lattner     unsigned InsertOffset = 0;
4058879e06dSChris Lattner     for (unsigned i = 0, e = VariableUses.size(); i != e; ++i) {
40692987fb3SAlexander Kornienko       std::string Value;
40792987fb3SAlexander Kornienko 
40892987fb3SAlexander Kornienko       if (VariableUses[i].first[0] == '@') {
40992987fb3SAlexander Kornienko         if (!EvaluateExpression(VariableUses[i].first, Value))
41092987fb3SAlexander Kornienko           return StringRef::npos;
41192987fb3SAlexander Kornienko       } else {
412e0ef65abSDaniel Dunbar         StringMap<StringRef>::iterator it =
413e0ef65abSDaniel Dunbar           VariableTable.find(VariableUses[i].first);
414e0ef65abSDaniel Dunbar         // If the variable is undefined, return an error.
415e0ef65abSDaniel Dunbar         if (it == VariableTable.end())
416e0ef65abSDaniel Dunbar           return StringRef::npos;
417e0ef65abSDaniel Dunbar 
4188879e06dSChris Lattner         // Look up the value and escape it so that we can plop it into the regex.
419e0ef65abSDaniel Dunbar         AddFixedStringToRegEx(it->second, Value);
42092987fb3SAlexander Kornienko       }
4218879e06dSChris Lattner 
4228879e06dSChris Lattner       // Plop it into the regex at the adjusted offset.
4238879e06dSChris Lattner       TmpStr.insert(TmpStr.begin()+VariableUses[i].second+InsertOffset,
4248879e06dSChris Lattner                     Value.begin(), Value.end());
4258879e06dSChris Lattner       InsertOffset += Value.size();
4268879e06dSChris Lattner     }
4278879e06dSChris Lattner 
4288879e06dSChris Lattner     // Match the newly constructed regex.
4298879e06dSChris Lattner     RegExToMatch = TmpStr;
4308879e06dSChris Lattner   }
4318879e06dSChris Lattner 
4328879e06dSChris Lattner 
433b16ab0c4SChris Lattner   SmallVector<StringRef, 4> MatchInfo;
4348879e06dSChris Lattner   if (!Regex(RegExToMatch, Regex::Newline).match(Buffer, &MatchInfo))
435f08d2db9SChris Lattner     return StringRef::npos;
436b16ab0c4SChris Lattner 
437b16ab0c4SChris Lattner   // Successful regex match.
438b16ab0c4SChris Lattner   assert(!MatchInfo.empty() && "Didn't get any match");
439b16ab0c4SChris Lattner   StringRef FullMatch = MatchInfo[0];
440b16ab0c4SChris Lattner 
4418879e06dSChris Lattner   // If this defines any variables, remember their values.
442e8b8f1bcSEli Bendersky   for (std::map<StringRef, unsigned>::const_iterator I = VariableDefs.begin(),
443e8b8f1bcSEli Bendersky                                                      E = VariableDefs.end();
444e8b8f1bcSEli Bendersky        I != E; ++I) {
445e8b8f1bcSEli Bendersky     assert(I->second < MatchInfo.size() && "Internal paren error");
446e8b8f1bcSEli Bendersky     VariableTable[I->first] = MatchInfo[I->second];
4470a4c44bdSChris Lattner   }
4480a4c44bdSChris Lattner 
449b16ab0c4SChris Lattner   MatchLen = FullMatch.size();
450b16ab0c4SChris Lattner   return FullMatch.data()-Buffer.data();
451f08d2db9SChris Lattner }
452f08d2db9SChris Lattner 
453fd29d886SDaniel Dunbar unsigned Pattern::ComputeMatchDistance(StringRef Buffer,
454fd29d886SDaniel Dunbar                               const StringMap<StringRef> &VariableTable) const {
455fd29d886SDaniel Dunbar   // Just compute the number of matching characters. For regular expressions, we
456fd29d886SDaniel Dunbar   // just compare against the regex itself and hope for the best.
457fd29d886SDaniel Dunbar   //
458fd29d886SDaniel Dunbar   // FIXME: One easy improvement here is have the regex lib generate a single
459fd29d886SDaniel Dunbar   // example regular expression which matches, and use that as the example
460fd29d886SDaniel Dunbar   // string.
461fd29d886SDaniel Dunbar   StringRef ExampleString(FixedStr);
462fd29d886SDaniel Dunbar   if (ExampleString.empty())
463fd29d886SDaniel Dunbar     ExampleString = RegExStr;
464fd29d886SDaniel Dunbar 
465e9aa36c8SDaniel Dunbar   // Only compare up to the first line in the buffer, or the string size.
466e9aa36c8SDaniel Dunbar   StringRef BufferPrefix = Buffer.substr(0, ExampleString.size());
467e9aa36c8SDaniel Dunbar   BufferPrefix = BufferPrefix.split('\n').first;
468e9aa36c8SDaniel Dunbar   return BufferPrefix.edit_distance(ExampleString);
469fd29d886SDaniel Dunbar }
470fd29d886SDaniel Dunbar 
471e0ef65abSDaniel Dunbar void Pattern::PrintFailureInfo(const SourceMgr &SM, StringRef Buffer,
472e0ef65abSDaniel Dunbar                                const StringMap<StringRef> &VariableTable) const{
473e0ef65abSDaniel Dunbar   // If this was a regular expression using variables, print the current
474e0ef65abSDaniel Dunbar   // variable values.
475e0ef65abSDaniel Dunbar   if (!VariableUses.empty()) {
476e0ef65abSDaniel Dunbar     for (unsigned i = 0, e = VariableUses.size(); i != e; ++i) {
477e0ef65abSDaniel Dunbar       SmallString<256> Msg;
478e0ef65abSDaniel Dunbar       raw_svector_ostream OS(Msg);
47992987fb3SAlexander Kornienko       StringRef Var = VariableUses[i].first;
48092987fb3SAlexander Kornienko       if (Var[0] == '@') {
48192987fb3SAlexander Kornienko         std::string Value;
48292987fb3SAlexander Kornienko         if (EvaluateExpression(Var, Value)) {
48392987fb3SAlexander Kornienko           OS << "with expression \"";
48492987fb3SAlexander Kornienko           OS.write_escaped(Var) << "\" equal to \"";
48592987fb3SAlexander Kornienko           OS.write_escaped(Value) << "\"";
48692987fb3SAlexander Kornienko         } else {
48792987fb3SAlexander Kornienko           OS << "uses incorrect expression \"";
48892987fb3SAlexander Kornienko           OS.write_escaped(Var) << "\"";
48992987fb3SAlexander Kornienko         }
49092987fb3SAlexander Kornienko       } else {
49192987fb3SAlexander Kornienko         StringMap<StringRef>::const_iterator it = VariableTable.find(Var);
492e0ef65abSDaniel Dunbar 
493e0ef65abSDaniel Dunbar         // Check for undefined variable references.
494e0ef65abSDaniel Dunbar         if (it == VariableTable.end()) {
495e0ef65abSDaniel Dunbar           OS << "uses undefined variable \"";
49692987fb3SAlexander Kornienko           OS.write_escaped(Var) << "\"";
497e0ef65abSDaniel Dunbar         } else {
498e0ef65abSDaniel Dunbar           OS << "with variable \"";
499e0ef65abSDaniel Dunbar           OS.write_escaped(Var) << "\" equal to \"";
500e0ef65abSDaniel Dunbar           OS.write_escaped(it->second) << "\"";
501e0ef65abSDaniel Dunbar         }
50292987fb3SAlexander Kornienko       }
503e0ef65abSDaniel Dunbar 
50403b80a40SChris Lattner       SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Note,
50503b80a40SChris Lattner                       OS.str());
506e0ef65abSDaniel Dunbar     }
507e0ef65abSDaniel Dunbar   }
508fd29d886SDaniel Dunbar 
509fd29d886SDaniel Dunbar   // Attempt to find the closest/best fuzzy match.  Usually an error happens
510fd29d886SDaniel Dunbar   // because some string in the output didn't exactly match. In these cases, we
511fd29d886SDaniel Dunbar   // would like to show the user a best guess at what "should have" matched, to
512fd29d886SDaniel Dunbar   // save them having to actually check the input manually.
513fd29d886SDaniel Dunbar   size_t NumLinesForward = 0;
514fd29d886SDaniel Dunbar   size_t Best = StringRef::npos;
515fd29d886SDaniel Dunbar   double BestQuality = 0;
516fd29d886SDaniel Dunbar 
517fd29d886SDaniel Dunbar   // Use an arbitrary 4k limit on how far we will search.
5182bf486ebSDan Gohman   for (size_t i = 0, e = std::min(size_t(4096), Buffer.size()); i != e; ++i) {
519fd29d886SDaniel Dunbar     if (Buffer[i] == '\n')
520fd29d886SDaniel Dunbar       ++NumLinesForward;
521fd29d886SDaniel Dunbar 
522df22bbf7SDan Gohman     // Patterns have leading whitespace stripped, so skip whitespace when
523df22bbf7SDan Gohman     // looking for something which looks like a pattern.
524df22bbf7SDan Gohman     if (Buffer[i] == ' ' || Buffer[i] == '\t')
525df22bbf7SDan Gohman       continue;
526df22bbf7SDan Gohman 
527fd29d886SDaniel Dunbar     // Compute the "quality" of this match as an arbitrary combination of the
528fd29d886SDaniel Dunbar     // match distance and the number of lines skipped to get to this match.
529fd29d886SDaniel Dunbar     unsigned Distance = ComputeMatchDistance(Buffer.substr(i), VariableTable);
530fd29d886SDaniel Dunbar     double Quality = Distance + (NumLinesForward / 100.);
531fd29d886SDaniel Dunbar 
532fd29d886SDaniel Dunbar     if (Quality < BestQuality || Best == StringRef::npos) {
533fd29d886SDaniel Dunbar       Best = i;
534fd29d886SDaniel Dunbar       BestQuality = Quality;
535fd29d886SDaniel Dunbar     }
536fd29d886SDaniel Dunbar   }
537fd29d886SDaniel Dunbar 
538fd29d886SDaniel Dunbar   // Print the "possible intended match here" line if we found something
539c069cc8eSDaniel Dunbar   // reasonable and not equal to what we showed in the "scanning from here"
540c069cc8eSDaniel Dunbar   // line.
541c069cc8eSDaniel Dunbar   if (Best && Best != StringRef::npos && BestQuality < 50) {
542fd29d886SDaniel Dunbar       SM.PrintMessage(SMLoc::getFromPointer(Buffer.data() + Best),
54303b80a40SChris Lattner                       SourceMgr::DK_Note, "possible intended match here");
544fd29d886SDaniel Dunbar 
545fd29d886SDaniel Dunbar     // FIXME: If we wanted to be really friendly we would show why the match
546fd29d886SDaniel Dunbar     // failed, as it can be hard to spot simple one character differences.
547fd29d886SDaniel Dunbar   }
548e0ef65abSDaniel Dunbar }
54974d50731SChris Lattner 
550061d2baaSEli Bendersky size_t Pattern::FindRegexVarEnd(StringRef Str) {
551061d2baaSEli Bendersky   // Offset keeps track of the current offset within the input Str
552061d2baaSEli Bendersky   size_t Offset = 0;
553061d2baaSEli Bendersky   // [...] Nesting depth
554061d2baaSEli Bendersky   size_t BracketDepth = 0;
555061d2baaSEli Bendersky 
556061d2baaSEli Bendersky   while (!Str.empty()) {
557061d2baaSEli Bendersky     if (Str.startswith("]]") && BracketDepth == 0)
558061d2baaSEli Bendersky       return Offset;
559061d2baaSEli Bendersky     if (Str[0] == '\\') {
560061d2baaSEli Bendersky       // Backslash escapes the next char within regexes, so skip them both.
561061d2baaSEli Bendersky       Str = Str.substr(2);
562061d2baaSEli Bendersky       Offset += 2;
563061d2baaSEli Bendersky     } else {
564061d2baaSEli Bendersky       switch (Str[0]) {
565061d2baaSEli Bendersky         default:
566061d2baaSEli Bendersky           break;
567061d2baaSEli Bendersky         case '[':
568061d2baaSEli Bendersky           BracketDepth++;
569061d2baaSEli Bendersky           break;
570061d2baaSEli Bendersky         case ']':
571061d2baaSEli Bendersky           assert(BracketDepth > 0 && "Invalid regex");
572061d2baaSEli Bendersky           BracketDepth--;
573061d2baaSEli Bendersky           break;
574061d2baaSEli Bendersky       }
575061d2baaSEli Bendersky       Str = Str.substr(1);
576061d2baaSEli Bendersky       Offset++;
577061d2baaSEli Bendersky     }
578061d2baaSEli Bendersky   }
579061d2baaSEli Bendersky 
580061d2baaSEli Bendersky   return StringRef::npos;
581061d2baaSEli Bendersky }
582061d2baaSEli Bendersky 
583061d2baaSEli Bendersky 
58474d50731SChris Lattner //===----------------------------------------------------------------------===//
58574d50731SChris Lattner // Check Strings.
58674d50731SChris Lattner //===----------------------------------------------------------------------===//
5873b40b445SChris Lattner 
5883b40b445SChris Lattner /// CheckString - This is a check that we found in the input file.
5893b40b445SChris Lattner struct CheckString {
5903b40b445SChris Lattner   /// Pat - The pattern to match.
5913b40b445SChris Lattner   Pattern Pat;
59226cccfe1SChris Lattner 
59326cccfe1SChris Lattner   /// Loc - The location in the match file that the check string was specified.
59426cccfe1SChris Lattner   SMLoc Loc;
59526cccfe1SChris Lattner 
596da108b4eSChris Lattner   /// IsCheckNext - This is true if this is a CHECK-NEXT: directive (as opposed
597da108b4eSChris Lattner   /// to a CHECK: directive.
598da108b4eSChris Lattner   bool IsCheckNext;
599da108b4eSChris Lattner 
600*f8bd2e5bSStephen Lin   /// IsCheckLabel - This is true if this is a CHECK-LABEL: directive (as
601*f8bd2e5bSStephen Lin   /// opposed to a CHECK: directive.
602*f8bd2e5bSStephen Lin   bool IsCheckLabel;
603*f8bd2e5bSStephen Lin 
60491a1b2c9SMichael Liao   /// DagNotStrings - These are all of the strings that are disallowed from
605236d2d5eSChris Lattner   /// occurring between this match string and the previous one (or start of
606236d2d5eSChris Lattner   /// file).
60791a1b2c9SMichael Liao   std::vector<Pattern> DagNotStrings;
608236d2d5eSChris Lattner 
609*f8bd2e5bSStephen Lin   CheckString(const Pattern &P, SMLoc L, bool isCheckNext, bool isCheckLabel)
610*f8bd2e5bSStephen Lin     : Pat(P), Loc(L), IsCheckNext(isCheckNext), IsCheckLabel(isCheckLabel) {}
611dcc7d48dSMichael Liao 
61291a1b2c9SMichael Liao   /// Check - Match check string and its "not strings" and/or "dag strings".
613*f8bd2e5bSStephen Lin   size_t Check(const SourceMgr &SM, StringRef Buffer, bool IsLabel,
614*f8bd2e5bSStephen Lin                size_t &MatchLen, StringMap<StringRef> &VariableTable) const;
615dcc7d48dSMichael Liao 
616dcc7d48dSMichael Liao   /// CheckNext - Verify there is a single line in the given buffer.
617dcc7d48dSMichael Liao   bool CheckNext(const SourceMgr &SM, StringRef Buffer) const;
618dcc7d48dSMichael Liao 
619dcc7d48dSMichael Liao   /// CheckNot - Verify there's no "not strings" in the given buffer.
620dcc7d48dSMichael Liao   bool CheckNot(const SourceMgr &SM, StringRef Buffer,
62191a1b2c9SMichael Liao                 const std::vector<const Pattern *> &NotStrings,
62291a1b2c9SMichael Liao                 StringMap<StringRef> &VariableTable) const;
62391a1b2c9SMichael Liao 
62491a1b2c9SMichael Liao   /// CheckDag - Match "dag strings" and their mixed "not strings".
62591a1b2c9SMichael Liao   size_t CheckDag(const SourceMgr &SM, StringRef Buffer,
62691a1b2c9SMichael Liao                   std::vector<const Pattern *> &NotStrings,
627dcc7d48dSMichael Liao                   StringMap<StringRef> &VariableTable) const;
62826cccfe1SChris Lattner };
62926cccfe1SChris Lattner 
6305ea04c38SGuy Benyei /// Canonicalize whitespaces in the input file. Line endings are replaced
6315ea04c38SGuy Benyei /// with UNIX-style '\n'.
6325ea04c38SGuy Benyei ///
6335ea04c38SGuy Benyei /// \param PreserveHorizontal Don't squash consecutive horizontal whitespace
6345ea04c38SGuy Benyei /// characters to a single space.
6355ea04c38SGuy Benyei static MemoryBuffer *CanonicalizeInputFile(MemoryBuffer *MB,
6365ea04c38SGuy Benyei                                            bool PreserveHorizontal) {
6370e45d24aSChris Lattner   SmallString<128> NewFile;
638a2f8fc5aSChris Lattner   NewFile.reserve(MB->getBufferSize());
639a2f8fc5aSChris Lattner 
640a2f8fc5aSChris Lattner   for (const char *Ptr = MB->getBufferStart(), *End = MB->getBufferEnd();
641a2f8fc5aSChris Lattner        Ptr != End; ++Ptr) {
642fd781bf0SNAKAMURA Takumi     // Eliminate trailing dosish \r.
643fd781bf0SNAKAMURA Takumi     if (Ptr <= End - 2 && Ptr[0] == '\r' && Ptr[1] == '\n') {
644fd781bf0SNAKAMURA Takumi       continue;
645fd781bf0SNAKAMURA Takumi     }
646fd781bf0SNAKAMURA Takumi 
6475ea04c38SGuy Benyei     // If current char is not a horizontal whitespace or if horizontal
6485ea04c38SGuy Benyei     // whitespace canonicalization is disabled, dump it to output as is.
6495ea04c38SGuy Benyei     if (PreserveHorizontal || (*Ptr != ' ' && *Ptr != '\t')) {
650a2f8fc5aSChris Lattner       NewFile.push_back(*Ptr);
651a2f8fc5aSChris Lattner       continue;
652a2f8fc5aSChris Lattner     }
653a2f8fc5aSChris Lattner 
654a2f8fc5aSChris Lattner     // Otherwise, add one space and advance over neighboring space.
655a2f8fc5aSChris Lattner     NewFile.push_back(' ');
656a2f8fc5aSChris Lattner     while (Ptr+1 != End &&
657a2f8fc5aSChris Lattner            (Ptr[1] == ' ' || Ptr[1] == '\t'))
658a2f8fc5aSChris Lattner       ++Ptr;
659a2f8fc5aSChris Lattner   }
660a2f8fc5aSChris Lattner 
661a2f8fc5aSChris Lattner   // Free the old buffer and return a new one.
662a2f8fc5aSChris Lattner   MemoryBuffer *MB2 =
6630e45d24aSChris Lattner     MemoryBuffer::getMemBufferCopy(NewFile.str(), MB->getBufferIdentifier());
664a2f8fc5aSChris Lattner 
665a2f8fc5aSChris Lattner   delete MB;
666a2f8fc5aSChris Lattner   return MB2;
667a2f8fc5aSChris Lattner }
668a2f8fc5aSChris Lattner 
669ee3c74fbSChris Lattner 
670ee3c74fbSChris Lattner /// ReadCheckFile - Read the check file, which specifies the sequence of
671ee3c74fbSChris Lattner /// expected strings.  The strings are added to the CheckStrings vector.
67243d50d4aSEli Bendersky /// Returns true in case of an error, false otherwise.
673ee3c74fbSChris Lattner static bool ReadCheckFile(SourceMgr &SM,
67426cccfe1SChris Lattner                           std::vector<CheckString> &CheckStrings) {
67539a0ffc3SMichael J. Spencer   OwningPtr<MemoryBuffer> File;
67639a0ffc3SMichael J. Spencer   if (error_code ec =
6778c811724SRafael Espindola         MemoryBuffer::getFileOrSTDIN(CheckFilename, File)) {
678ee3c74fbSChris Lattner     errs() << "Could not open check file '" << CheckFilename << "': "
6797b6fef82SMichael J. Spencer            << ec.message() << '\n';
680ee3c74fbSChris Lattner     return true;
681ee3c74fbSChris Lattner   }
682a2f8fc5aSChris Lattner 
683a2f8fc5aSChris Lattner   // If we want to canonicalize whitespace, strip excess whitespace from the
6845ea04c38SGuy Benyei   // buffer containing the CHECK lines. Remove DOS style line endings.
685e963d660SBenjamin Kramer   MemoryBuffer *F =
686e963d660SBenjamin Kramer     CanonicalizeInputFile(File.take(), NoCanonicalizeWhiteSpace);
687a2f8fc5aSChris Lattner 
688ee3c74fbSChris Lattner   SM.AddNewSourceBuffer(F, SMLoc());
689ee3c74fbSChris Lattner 
69010f10cedSChris Lattner   // Find all instances of CheckPrefix followed by : in the file.
691caa5fc0cSChris Lattner   StringRef Buffer = F->getBuffer();
69291a1b2c9SMichael Liao   std::vector<Pattern> DagNotMatches;
693236d2d5eSChris Lattner 
69443d50d4aSEli Bendersky   // LineNumber keeps track of the line on which CheckPrefix instances are
69543d50d4aSEli Bendersky   // found.
69692987fb3SAlexander Kornienko   unsigned LineNumber = 1;
69792987fb3SAlexander Kornienko 
698ee3c74fbSChris Lattner   while (1) {
699ee3c74fbSChris Lattner     // See if Prefix occurs in the memory buffer.
70092987fb3SAlexander Kornienko     size_t PrefixLoc = Buffer.find(CheckPrefix);
701ee3c74fbSChris Lattner     // If we didn't find a match, we're done.
70292987fb3SAlexander Kornienko     if (PrefixLoc == StringRef::npos)
703ee3c74fbSChris Lattner       break;
704ee3c74fbSChris Lattner 
70592987fb3SAlexander Kornienko     LineNumber += Buffer.substr(0, PrefixLoc).count('\n');
70692987fb3SAlexander Kornienko 
70792987fb3SAlexander Kornienko     Buffer = Buffer.substr(PrefixLoc);
70892987fb3SAlexander Kornienko 
709caa5fc0cSChris Lattner     const char *CheckPrefixStart = Buffer.data();
710da108b4eSChris Lattner 
711da108b4eSChris Lattner     // When we find a check prefix, keep track of whether we find CHECK: or
712da108b4eSChris Lattner     // CHECK-NEXT:
713*f8bd2e5bSStephen Lin     bool IsCheckNext = false, IsCheckNot = false, IsCheckDag = false,
714*f8bd2e5bSStephen Lin          IsCheckLabel = false;
715da108b4eSChris Lattner 
71610f10cedSChris Lattner     // Verify that the : is present after the prefix.
717caa5fc0cSChris Lattner     if (Buffer[CheckPrefix.size()] == ':') {
718caa5fc0cSChris Lattner       Buffer = Buffer.substr(CheckPrefix.size()+1);
719caa5fc0cSChris Lattner     } else if (Buffer.size() > CheckPrefix.size()+6 &&
720caa5fc0cSChris Lattner                memcmp(Buffer.data()+CheckPrefix.size(), "-NEXT:", 6) == 0) {
721ef878a83SBenjamin Kramer       Buffer = Buffer.substr(CheckPrefix.size()+6);
722da108b4eSChris Lattner       IsCheckNext = true;
723236d2d5eSChris Lattner     } else if (Buffer.size() > CheckPrefix.size()+5 &&
724236d2d5eSChris Lattner                memcmp(Buffer.data()+CheckPrefix.size(), "-NOT:", 5) == 0) {
725ef878a83SBenjamin Kramer       Buffer = Buffer.substr(CheckPrefix.size()+5);
726236d2d5eSChris Lattner       IsCheckNot = true;
72791a1b2c9SMichael Liao     } else if (Buffer.size() > CheckPrefix.size()+5 &&
72891a1b2c9SMichael Liao                memcmp(Buffer.data()+CheckPrefix.size(), "-DAG:", 5) == 0) {
72991a1b2c9SMichael Liao       Buffer = Buffer.substr(CheckPrefix.size()+5);
73091a1b2c9SMichael Liao       IsCheckDag = true;
731*f8bd2e5bSStephen Lin     } else if (Buffer.size() > CheckPrefix.size()+7 &&
732*f8bd2e5bSStephen Lin                memcmp(Buffer.data()+CheckPrefix.size(), "-LABEL:", 7) == 0) {
733*f8bd2e5bSStephen Lin       Buffer = Buffer.substr(CheckPrefix.size()+7);
734*f8bd2e5bSStephen Lin       IsCheckLabel = true;
735da108b4eSChris Lattner     } else {
736caa5fc0cSChris Lattner       Buffer = Buffer.substr(1);
73710f10cedSChris Lattner       continue;
73810f10cedSChris Lattner     }
73910f10cedSChris Lattner 
740ee3c74fbSChris Lattner     // Okay, we found the prefix, yay.  Remember the rest of the line, but
741ee3c74fbSChris Lattner     // ignore leading and trailing whitespace.
742236d2d5eSChris Lattner     Buffer = Buffer.substr(Buffer.find_first_not_of(" \t"));
743ee3c74fbSChris Lattner 
744ee3c74fbSChris Lattner     // Scan ahead to the end of line.
745caa5fc0cSChris Lattner     size_t EOL = Buffer.find_first_of("\n\r");
746ee3c74fbSChris Lattner 
747838fb09aSDan Gohman     // Remember the location of the start of the pattern, for diagnostics.
748838fb09aSDan Gohman     SMLoc PatternLoc = SMLoc::getFromPointer(Buffer.data());
749838fb09aSDan Gohman 
75074d50731SChris Lattner     // Parse the pattern.
75174d50731SChris Lattner     Pattern P;
75292987fb3SAlexander Kornienko     if (P.ParsePattern(Buffer.substr(0, EOL), SM, LineNumber))
753ee3c74fbSChris Lattner       return true;
754ee3c74fbSChris Lattner 
755*f8bd2e5bSStephen Lin     // Verify that CHECK-LABEL lines do not define or use variables
756*f8bd2e5bSStephen Lin     if (IsCheckLabel && P.hasVariable()) {
757*f8bd2e5bSStephen Lin       SM.PrintMessage(SMLoc::getFromPointer(CheckPrefixStart),
758*f8bd2e5bSStephen Lin                       SourceMgr::DK_Error,
759*f8bd2e5bSStephen Lin                       "found '"+CheckPrefix+"-LABEL:' with variable definition"
760*f8bd2e5bSStephen Lin                       " or use'");
761*f8bd2e5bSStephen Lin       return true;
762*f8bd2e5bSStephen Lin     }
763*f8bd2e5bSStephen Lin 
76491a1b2c9SMichael Liao     P.setMatchNot(IsCheckNot);
76591a1b2c9SMichael Liao     P.setMatchDag(IsCheckDag);
76691a1b2c9SMichael Liao 
767236d2d5eSChris Lattner     Buffer = Buffer.substr(EOL);
76874d50731SChris Lattner 
769da108b4eSChris Lattner     // Verify that CHECK-NEXT lines have at least one CHECK line before them.
770da108b4eSChris Lattner     if (IsCheckNext && CheckStrings.empty()) {
771da108b4eSChris Lattner       SM.PrintMessage(SMLoc::getFromPointer(CheckPrefixStart),
77203b80a40SChris Lattner                       SourceMgr::DK_Error,
773da108b4eSChris Lattner                       "found '"+CheckPrefix+"-NEXT:' without previous '"+
77403b80a40SChris Lattner                       CheckPrefix+ ": line");
775da108b4eSChris Lattner       return true;
776da108b4eSChris Lattner     }
777da108b4eSChris Lattner 
77891a1b2c9SMichael Liao     // Handle CHECK-DAG/-NOT.
77991a1b2c9SMichael Liao     if (IsCheckDag || IsCheckNot) {
78091a1b2c9SMichael Liao       DagNotMatches.push_back(P);
78174d50731SChris Lattner       continue;
78274d50731SChris Lattner     }
78374d50731SChris Lattner 
784ee3c74fbSChris Lattner     // Okay, add the string we captured to the output vector and move on.
7853b40b445SChris Lattner     CheckStrings.push_back(CheckString(P,
786838fb09aSDan Gohman                                        PatternLoc,
787*f8bd2e5bSStephen Lin                                        IsCheckNext,
788*f8bd2e5bSStephen Lin                                        IsCheckLabel));
78991a1b2c9SMichael Liao     std::swap(DagNotMatches, CheckStrings.back().DagNotStrings);
790ee3c74fbSChris Lattner   }
791ee3c74fbSChris Lattner 
79291a1b2c9SMichael Liao   // Add an EOF pattern for any trailing CHECK-DAG/-NOTs.
79391a1b2c9SMichael Liao   if (!DagNotMatches.empty()) {
794eba55822SJakob Stoklund Olesen     CheckStrings.push_back(CheckString(Pattern(true),
795eba55822SJakob Stoklund Olesen                                        SMLoc::getFromPointer(Buffer.data()),
796*f8bd2e5bSStephen Lin                                        false,
797eba55822SJakob Stoklund Olesen                                        false));
79891a1b2c9SMichael Liao     std::swap(DagNotMatches, CheckStrings.back().DagNotStrings);
799eba55822SJakob Stoklund Olesen   }
800eba55822SJakob Stoklund Olesen 
801ee3c74fbSChris Lattner   if (CheckStrings.empty()) {
80210f10cedSChris Lattner     errs() << "error: no check strings found with prefix '" << CheckPrefix
80310f10cedSChris Lattner            << ":'\n";
804ee3c74fbSChris Lattner     return true;
805ee3c74fbSChris Lattner   }
806ee3c74fbSChris Lattner 
807ee3c74fbSChris Lattner   return false;
808ee3c74fbSChris Lattner }
809ee3c74fbSChris Lattner 
81091a1b2c9SMichael Liao static void PrintCheckFailed(const SourceMgr &SM, const SMLoc &Loc,
81191a1b2c9SMichael Liao                              const Pattern &Pat, StringRef Buffer,
812e0ef65abSDaniel Dunbar                              StringMap<StringRef> &VariableTable) {
813da108b4eSChris Lattner   // Otherwise, we have an error, emit an error message.
81491a1b2c9SMichael Liao   SM.PrintMessage(Loc, SourceMgr::DK_Error,
81503b80a40SChris Lattner                   "expected string not found in input");
816da108b4eSChris Lattner 
817da108b4eSChris Lattner   // Print the "scanning from here" line.  If the current position is at the
818da108b4eSChris Lattner   // end of a line, advance to the start of the next line.
819caa5fc0cSChris Lattner   Buffer = Buffer.substr(Buffer.find_first_not_of(" \t\n\r"));
820da108b4eSChris Lattner 
82103b80a40SChris Lattner   SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Note,
82203b80a40SChris Lattner                   "scanning from here");
823e0ef65abSDaniel Dunbar 
824e0ef65abSDaniel Dunbar   // Allow the pattern to print additional information if desired.
82591a1b2c9SMichael Liao   Pat.PrintFailureInfo(SM, Buffer, VariableTable);
82691a1b2c9SMichael Liao }
82791a1b2c9SMichael Liao 
82891a1b2c9SMichael Liao static void PrintCheckFailed(const SourceMgr &SM, const CheckString &CheckStr,
82991a1b2c9SMichael Liao                              StringRef Buffer,
83091a1b2c9SMichael Liao                              StringMap<StringRef> &VariableTable) {
83191a1b2c9SMichael Liao   PrintCheckFailed(SM, CheckStr.Loc, CheckStr.Pat, Buffer, VariableTable);
832da108b4eSChris Lattner }
833da108b4eSChris Lattner 
83437183584SChris Lattner /// CountNumNewlinesBetween - Count the number of newlines in the specified
83537183584SChris Lattner /// range.
83637183584SChris Lattner static unsigned CountNumNewlinesBetween(StringRef Range) {
837da108b4eSChris Lattner   unsigned NumNewLines = 0;
83837183584SChris Lattner   while (1) {
839da108b4eSChris Lattner     // Scan for newline.
84037183584SChris Lattner     Range = Range.substr(Range.find_first_of("\n\r"));
84137183584SChris Lattner     if (Range.empty()) return NumNewLines;
842da108b4eSChris Lattner 
843da108b4eSChris Lattner     ++NumNewLines;
844da108b4eSChris Lattner 
845da108b4eSChris Lattner     // Handle \n\r and \r\n as a single newline.
84637183584SChris Lattner     if (Range.size() > 1 &&
84737183584SChris Lattner         (Range[1] == '\n' || Range[1] == '\r') &&
84837183584SChris Lattner         (Range[0] != Range[1]))
84937183584SChris Lattner       Range = Range.substr(1);
85037183584SChris Lattner     Range = Range.substr(1);
851da108b4eSChris Lattner   }
852da108b4eSChris Lattner }
853da108b4eSChris Lattner 
854dcc7d48dSMichael Liao size_t CheckString::Check(const SourceMgr &SM, StringRef Buffer,
855*f8bd2e5bSStephen Lin                           bool IsLabel, size_t &MatchLen,
856dcc7d48dSMichael Liao                           StringMap<StringRef> &VariableTable) const {
85791a1b2c9SMichael Liao   size_t LastPos = 0;
85891a1b2c9SMichael Liao   std::vector<const Pattern *> NotStrings;
85991a1b2c9SMichael Liao 
860*f8bd2e5bSStephen Lin   if (!IsLabel) {
86191a1b2c9SMichael Liao     // Match "dag strings" (with mixed "not strings" if any).
86291a1b2c9SMichael Liao     LastPos = CheckDag(SM, Buffer, NotStrings, VariableTable);
86391a1b2c9SMichael Liao     if (LastPos == StringRef::npos)
86491a1b2c9SMichael Liao       return StringRef::npos;
865*f8bd2e5bSStephen Lin   }
86691a1b2c9SMichael Liao 
86791a1b2c9SMichael Liao   // Match itself from the last position after matching CHECK-DAG.
86891a1b2c9SMichael Liao   StringRef MatchBuffer = Buffer.substr(LastPos);
86991a1b2c9SMichael Liao   size_t MatchPos = Pat.Match(MatchBuffer, MatchLen, VariableTable);
870dcc7d48dSMichael Liao   if (MatchPos == StringRef::npos) {
87191a1b2c9SMichael Liao     PrintCheckFailed(SM, *this, MatchBuffer, VariableTable);
872dcc7d48dSMichael Liao     return StringRef::npos;
873dcc7d48dSMichael Liao   }
87491a1b2c9SMichael Liao   MatchPos += LastPos;
875dcc7d48dSMichael Liao 
876*f8bd2e5bSStephen Lin   if (!IsLabel) {
87791a1b2c9SMichael Liao     StringRef SkippedRegion = Buffer.substr(LastPos, MatchPos);
878dcc7d48dSMichael Liao 
879dcc7d48dSMichael Liao     // If this check is a "CHECK-NEXT", verify that the previous match was on
880dcc7d48dSMichael Liao     // the previous line (i.e. that there is one newline between them).
881dcc7d48dSMichael Liao     if (CheckNext(SM, SkippedRegion))
882dcc7d48dSMichael Liao       return StringRef::npos;
883dcc7d48dSMichael Liao 
884dcc7d48dSMichael Liao     // If this match had "not strings", verify that they don't exist in the
885dcc7d48dSMichael Liao     // skipped region.
88691a1b2c9SMichael Liao     if (CheckNot(SM, SkippedRegion, NotStrings, VariableTable))
887dcc7d48dSMichael Liao       return StringRef::npos;
888*f8bd2e5bSStephen Lin   }
889dcc7d48dSMichael Liao 
890dcc7d48dSMichael Liao   return MatchPos;
891dcc7d48dSMichael Liao }
892dcc7d48dSMichael Liao 
893dcc7d48dSMichael Liao bool CheckString::CheckNext(const SourceMgr &SM, StringRef Buffer) const {
894dcc7d48dSMichael Liao   if (!IsCheckNext)
895dcc7d48dSMichael Liao     return false;
896dcc7d48dSMichael Liao 
897dcc7d48dSMichael Liao   // Count the number of newlines between the previous match and this one.
898dcc7d48dSMichael Liao   assert(Buffer.data() !=
899dcc7d48dSMichael Liao          SM.getMemoryBuffer(
900dcc7d48dSMichael Liao            SM.FindBufferContainingLoc(
901dcc7d48dSMichael Liao              SMLoc::getFromPointer(Buffer.data())))->getBufferStart() &&
902dcc7d48dSMichael Liao          "CHECK-NEXT can't be the first check in a file");
903dcc7d48dSMichael Liao 
904dcc7d48dSMichael Liao   unsigned NumNewLines = CountNumNewlinesBetween(Buffer);
905dcc7d48dSMichael Liao 
906dcc7d48dSMichael Liao   if (NumNewLines == 0) {
907dcc7d48dSMichael Liao     SM.PrintMessage(Loc, SourceMgr::DK_Error, CheckPrefix+
908dcc7d48dSMichael Liao                     "-NEXT: is on the same line as previous match");
909dcc7d48dSMichael Liao     SM.PrintMessage(SMLoc::getFromPointer(Buffer.end()),
910dcc7d48dSMichael Liao                     SourceMgr::DK_Note, "'next' match was here");
911dcc7d48dSMichael Liao     SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Note,
912dcc7d48dSMichael Liao                     "previous match ended here");
913dcc7d48dSMichael Liao     return true;
914dcc7d48dSMichael Liao   }
915dcc7d48dSMichael Liao 
916dcc7d48dSMichael Liao   if (NumNewLines != 1) {
917dcc7d48dSMichael Liao     SM.PrintMessage(Loc, SourceMgr::DK_Error, CheckPrefix+
918dcc7d48dSMichael Liao                     "-NEXT: is not on the line after the previous match");
919dcc7d48dSMichael Liao     SM.PrintMessage(SMLoc::getFromPointer(Buffer.end()),
920dcc7d48dSMichael Liao                     SourceMgr::DK_Note, "'next' match was here");
921dcc7d48dSMichael Liao     SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Note,
922dcc7d48dSMichael Liao                     "previous match ended here");
923dcc7d48dSMichael Liao     return true;
924dcc7d48dSMichael Liao   }
925dcc7d48dSMichael Liao 
926dcc7d48dSMichael Liao   return false;
927dcc7d48dSMichael Liao }
928dcc7d48dSMichael Liao 
929dcc7d48dSMichael Liao bool CheckString::CheckNot(const SourceMgr &SM, StringRef Buffer,
93091a1b2c9SMichael Liao                            const std::vector<const Pattern *> &NotStrings,
931dcc7d48dSMichael Liao                            StringMap<StringRef> &VariableTable) const {
932dcc7d48dSMichael Liao   for (unsigned ChunkNo = 0, e = NotStrings.size();
933dcc7d48dSMichael Liao        ChunkNo != e; ++ChunkNo) {
93491a1b2c9SMichael Liao     const Pattern *Pat = NotStrings[ChunkNo];
93591a1b2c9SMichael Liao     assert(Pat->getMatchNot() && "Expect CHECK-NOT!");
93691a1b2c9SMichael Liao 
937dcc7d48dSMichael Liao     size_t MatchLen = 0;
93891a1b2c9SMichael Liao     size_t Pos = Pat->Match(Buffer, MatchLen, VariableTable);
939dcc7d48dSMichael Liao 
940dcc7d48dSMichael Liao     if (Pos == StringRef::npos) continue;
941dcc7d48dSMichael Liao 
942dcc7d48dSMichael Liao     SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()+Pos),
943dcc7d48dSMichael Liao                     SourceMgr::DK_Error,
944dcc7d48dSMichael Liao                     CheckPrefix+"-NOT: string occurred!");
94591a1b2c9SMichael Liao     SM.PrintMessage(Pat->getLoc(), SourceMgr::DK_Note,
946dcc7d48dSMichael Liao                     CheckPrefix+"-NOT: pattern specified here");
947dcc7d48dSMichael Liao     return true;
948dcc7d48dSMichael Liao   }
949dcc7d48dSMichael Liao 
950dcc7d48dSMichael Liao   return false;
951dcc7d48dSMichael Liao }
952dcc7d48dSMichael Liao 
95391a1b2c9SMichael Liao size_t CheckString::CheckDag(const SourceMgr &SM, StringRef Buffer,
95491a1b2c9SMichael Liao                              std::vector<const Pattern *> &NotStrings,
95591a1b2c9SMichael Liao                              StringMap<StringRef> &VariableTable) const {
95691a1b2c9SMichael Liao   if (DagNotStrings.empty())
95791a1b2c9SMichael Liao     return 0;
95891a1b2c9SMichael Liao 
95991a1b2c9SMichael Liao   size_t LastPos = 0;
96091a1b2c9SMichael Liao   size_t StartPos = LastPos;
96191a1b2c9SMichael Liao 
96291a1b2c9SMichael Liao   for (unsigned ChunkNo = 0, e = DagNotStrings.size();
96391a1b2c9SMichael Liao        ChunkNo != e; ++ChunkNo) {
96491a1b2c9SMichael Liao     const Pattern &Pat = DagNotStrings[ChunkNo];
96591a1b2c9SMichael Liao 
96691a1b2c9SMichael Liao     assert((Pat.getMatchDag() ^ Pat.getMatchNot()) &&
96791a1b2c9SMichael Liao            "Invalid CHECK-DAG or CHECK-NOT!");
96891a1b2c9SMichael Liao 
96991a1b2c9SMichael Liao     if (Pat.getMatchNot()) {
97091a1b2c9SMichael Liao       NotStrings.push_back(&Pat);
97191a1b2c9SMichael Liao       continue;
97291a1b2c9SMichael Liao     }
97391a1b2c9SMichael Liao 
97491a1b2c9SMichael Liao     assert(Pat.getMatchDag() && "Expect CHECK-DAG!");
97591a1b2c9SMichael Liao 
97691a1b2c9SMichael Liao     size_t MatchLen = 0, MatchPos;
97791a1b2c9SMichael Liao 
97891a1b2c9SMichael Liao     // CHECK-DAG always matches from the start.
97991a1b2c9SMichael Liao     StringRef MatchBuffer = Buffer.substr(StartPos);
98091a1b2c9SMichael Liao     MatchPos = Pat.Match(MatchBuffer, MatchLen, VariableTable);
98191a1b2c9SMichael Liao     // With a group of CHECK-DAGs, a single mismatching means the match on
98291a1b2c9SMichael Liao     // that group of CHECK-DAGs fails immediately.
98391a1b2c9SMichael Liao     if (MatchPos == StringRef::npos) {
98491a1b2c9SMichael Liao       PrintCheckFailed(SM, Pat.getLoc(), Pat, MatchBuffer, VariableTable);
98591a1b2c9SMichael Liao       return StringRef::npos;
98691a1b2c9SMichael Liao     }
98791a1b2c9SMichael Liao     // Re-calc it as the offset relative to the start of the original string.
98891a1b2c9SMichael Liao     MatchPos += StartPos;
98991a1b2c9SMichael Liao 
99091a1b2c9SMichael Liao     if (!NotStrings.empty()) {
99191a1b2c9SMichael Liao       if (MatchPos < LastPos) {
99291a1b2c9SMichael Liao         // Reordered?
99391a1b2c9SMichael Liao         SM.PrintMessage(SMLoc::getFromPointer(Buffer.data() + MatchPos),
99491a1b2c9SMichael Liao                         SourceMgr::DK_Error,
99591a1b2c9SMichael Liao                         CheckPrefix+"-DAG: found a match of CHECK-DAG"
99691a1b2c9SMichael Liao                         " reordering across a CHECK-NOT");
99791a1b2c9SMichael Liao         SM.PrintMessage(SMLoc::getFromPointer(Buffer.data() + LastPos),
99891a1b2c9SMichael Liao                         SourceMgr::DK_Note,
99991a1b2c9SMichael Liao                         CheckPrefix+"-DAG: the farthest match of CHECK-DAG"
100091a1b2c9SMichael Liao                         " is found here");
100191a1b2c9SMichael Liao         SM.PrintMessage(NotStrings[0]->getLoc(), SourceMgr::DK_Note,
100291a1b2c9SMichael Liao                         CheckPrefix+"-NOT: the crossed pattern specified"
100391a1b2c9SMichael Liao                         " here");
100491a1b2c9SMichael Liao         SM.PrintMessage(Pat.getLoc(), SourceMgr::DK_Note,
100591a1b2c9SMichael Liao                         CheckPrefix+"-DAG: the reordered pattern specified"
100691a1b2c9SMichael Liao                         " here");
100791a1b2c9SMichael Liao         return StringRef::npos;
100891a1b2c9SMichael Liao       }
100991a1b2c9SMichael Liao       // All subsequent CHECK-DAGs should be matched from the farthest
101091a1b2c9SMichael Liao       // position of all precedent CHECK-DAGs (including this one.)
101191a1b2c9SMichael Liao       StartPos = LastPos;
101291a1b2c9SMichael Liao       // If there's CHECK-NOTs between two CHECK-DAGs or from CHECK to
101391a1b2c9SMichael Liao       // CHECK-DAG, verify that there's no 'not' strings occurred in that
101491a1b2c9SMichael Liao       // region.
101591a1b2c9SMichael Liao       StringRef SkippedRegion = Buffer.substr(LastPos, MatchPos);
101691a1b2c9SMichael Liao       size_t Pos = CheckNot(SM, SkippedRegion, NotStrings, VariableTable);
101791a1b2c9SMichael Liao       if (Pos != StringRef::npos)
101891a1b2c9SMichael Liao         return StringRef::npos;
101991a1b2c9SMichael Liao       // Clear "not strings".
102091a1b2c9SMichael Liao       NotStrings.clear();
102191a1b2c9SMichael Liao     }
102291a1b2c9SMichael Liao 
102391a1b2c9SMichael Liao     // Update the last position with CHECK-DAG matches.
102491a1b2c9SMichael Liao     LastPos = std::max(MatchPos + MatchLen, LastPos);
102591a1b2c9SMichael Liao   }
102691a1b2c9SMichael Liao 
102791a1b2c9SMichael Liao   return LastPos;
102891a1b2c9SMichael Liao }
102991a1b2c9SMichael Liao 
1030ee3c74fbSChris Lattner int main(int argc, char **argv) {
1031ee3c74fbSChris Lattner   sys::PrintStackTraceOnErrorSignal();
1032ee3c74fbSChris Lattner   PrettyStackTraceProgram X(argc, argv);
1033ee3c74fbSChris Lattner   cl::ParseCommandLineOptions(argc, argv);
1034ee3c74fbSChris Lattner 
1035ee3c74fbSChris Lattner   SourceMgr SM;
1036ee3c74fbSChris Lattner 
1037ee3c74fbSChris Lattner   // Read the expected strings from the check file.
103826cccfe1SChris Lattner   std::vector<CheckString> CheckStrings;
1039ee3c74fbSChris Lattner   if (ReadCheckFile(SM, CheckStrings))
1040ee3c74fbSChris Lattner     return 2;
1041ee3c74fbSChris Lattner 
1042ee3c74fbSChris Lattner   // Open the file to check and add it to SourceMgr.
104339a0ffc3SMichael J. Spencer   OwningPtr<MemoryBuffer> File;
104439a0ffc3SMichael J. Spencer   if (error_code ec =
10458c811724SRafael Espindola         MemoryBuffer::getFileOrSTDIN(InputFilename, File)) {
1046ee3c74fbSChris Lattner     errs() << "Could not open input file '" << InputFilename << "': "
10477b6fef82SMichael J. Spencer            << ec.message() << '\n';
10488e1c6477SEli Bendersky     return 2;
1049ee3c74fbSChris Lattner   }
10502c3e5cdfSChris Lattner 
1051e963d660SBenjamin Kramer   if (File->getBufferSize() == 0) {
1052b692bed7SChris Lattner     errs() << "FileCheck error: '" << InputFilename << "' is empty.\n";
10538e1c6477SEli Bendersky     return 2;
1054b692bed7SChris Lattner   }
1055b692bed7SChris Lattner 
10562c3e5cdfSChris Lattner   // Remove duplicate spaces in the input file if requested.
10575ea04c38SGuy Benyei   // Remove DOS style line endings.
1058e963d660SBenjamin Kramer   MemoryBuffer *F =
1059e963d660SBenjamin Kramer     CanonicalizeInputFile(File.take(), NoCanonicalizeWhiteSpace);
10602c3e5cdfSChris Lattner 
1061ee3c74fbSChris Lattner   SM.AddNewSourceBuffer(F, SMLoc());
1062ee3c74fbSChris Lattner 
10638879e06dSChris Lattner   /// VariableTable - This holds all the current filecheck variables.
10648879e06dSChris Lattner   StringMap<StringRef> VariableTable;
10658879e06dSChris Lattner 
1066ee3c74fbSChris Lattner   // Check that we have all of the expected strings, in order, in the input
1067ee3c74fbSChris Lattner   // file.
1068caa5fc0cSChris Lattner   StringRef Buffer = F->getBuffer();
1069ee3c74fbSChris Lattner 
1070*f8bd2e5bSStephen Lin   bool hasError = false;
1071ee3c74fbSChris Lattner 
1072*f8bd2e5bSStephen Lin   unsigned i = 0, j = 0, e = CheckStrings.size();
1073ee3c74fbSChris Lattner 
1074*f8bd2e5bSStephen Lin   while (true) {
1075*f8bd2e5bSStephen Lin     StringRef CheckRegion;
1076*f8bd2e5bSStephen Lin     if (j == e) {
1077*f8bd2e5bSStephen Lin       CheckRegion = Buffer;
1078*f8bd2e5bSStephen Lin     } else {
1079*f8bd2e5bSStephen Lin       const CheckString &CheckLabelStr = CheckStrings[j];
1080*f8bd2e5bSStephen Lin       if (!CheckLabelStr.IsCheckLabel) {
1081*f8bd2e5bSStephen Lin         ++j;
1082*f8bd2e5bSStephen Lin         continue;
1083da108b4eSChris Lattner       }
1084da108b4eSChris Lattner 
1085*f8bd2e5bSStephen Lin       // Scan to next CHECK-LABEL match, ignoring CHECK-NOT and CHECK-DAG
1086*f8bd2e5bSStephen Lin       size_t MatchLabelLen = 0;
1087*f8bd2e5bSStephen Lin       size_t MatchLabelPos = CheckLabelStr.Check(SM, Buffer, true,
1088*f8bd2e5bSStephen Lin                                                  MatchLabelLen, VariableTable);
1089*f8bd2e5bSStephen Lin       if (MatchLabelPos == StringRef::npos) {
1090*f8bd2e5bSStephen Lin         hasError = true;
1091*f8bd2e5bSStephen Lin         break;
1092*f8bd2e5bSStephen Lin       }
1093*f8bd2e5bSStephen Lin 
1094*f8bd2e5bSStephen Lin       CheckRegion = Buffer.substr(0, MatchLabelPos + MatchLabelLen);
1095*f8bd2e5bSStephen Lin       Buffer = Buffer.substr(MatchLabelPos + MatchLabelLen);
1096*f8bd2e5bSStephen Lin       ++j;
1097*f8bd2e5bSStephen Lin     }
1098*f8bd2e5bSStephen Lin 
1099*f8bd2e5bSStephen Lin     for ( ; i != j; ++i) {
1100*f8bd2e5bSStephen Lin       const CheckString &CheckStr = CheckStrings[i];
1101*f8bd2e5bSStephen Lin 
1102*f8bd2e5bSStephen Lin       // Check each string within the scanned region, including a second check
1103*f8bd2e5bSStephen Lin       // of any final CHECK-LABEL (to verify CHECK-NOT and CHECK-DAG)
1104*f8bd2e5bSStephen Lin       size_t MatchLen = 0;
1105*f8bd2e5bSStephen Lin       size_t MatchPos = CheckStr.Check(SM, CheckRegion, false, MatchLen,
1106*f8bd2e5bSStephen Lin                                        VariableTable);
1107*f8bd2e5bSStephen Lin 
1108*f8bd2e5bSStephen Lin       if (MatchPos == StringRef::npos) {
1109*f8bd2e5bSStephen Lin         hasError = true;
1110*f8bd2e5bSStephen Lin         i = j;
1111*f8bd2e5bSStephen Lin         break;
1112*f8bd2e5bSStephen Lin       }
1113*f8bd2e5bSStephen Lin 
1114*f8bd2e5bSStephen Lin       CheckRegion = CheckRegion.substr(MatchPos + MatchLen);
1115*f8bd2e5bSStephen Lin     }
1116*f8bd2e5bSStephen Lin 
1117*f8bd2e5bSStephen Lin     if (j == e)
1118*f8bd2e5bSStephen Lin       break;
1119*f8bd2e5bSStephen Lin   }
1120*f8bd2e5bSStephen Lin 
1121*f8bd2e5bSStephen Lin   return hasError ? 1 : 0;
1122ee3c74fbSChris Lattner }
1123